AWS was a loss machine for 9 years though. They scaled a profitable retail business, invested in tech while folks would point out that they weren't profitable but they stuck it out for a decade. There was even a 2 year window when AWS lost money but Amazon as a whole was profitable, and AWS didn't overtake Amazon retail until after 2016.
I feel like people try to shoehorn anything tech into a market-based bucket - it gets too revisionist imho. I mean, you have all the telecoms and server vendors in the 90's experimenting with cloud computing, leasing servers to businesses, laying the infrastructure for broadband, any of them could have made cloud happen - and yet it's freaking Amazon who figures out how to monetize containerization and commodity x86 tech to define the cloud? Remember what streaming quality was like before Netflix pivoted away from shipping DVD's (to much uproar, by the way. Comical in hindsight). Remember the first time you took Uber? The first time you picked up an iPhone or spun the iPod dial? Snowflake's not going to be the next Cloudera because their product is vastly superior at this stage compared to what Cloudera's was (and I say that as someone who implemented more than $100M in Cloudera over the past 5 years).
I mean, anyone can Tivo themselves by just being absolute idiots in business, but in this recent era of essentially free money just snap-judging tech on profitability and scoffing at high valuations isn't that insightful. Those are the folks that mocked the Facebook IPO and reveled in the short-term drop. Nobody uses Myspace anymore and it's not because of marketing. Ask folks in AI if investing in a "gaming company that isn't relevant in the business world" was a good move 3 years ago with NVIDIA...